Applications are made up of a large number of instructions and data. Instructions operate on data which is fetched in a cache and memory and is always unencrypted. Scaled-out, distributed applications are made up of a large number of application instances. These application instances have their own data in the cache and memory of the processor on which these applications run. A large number of such application instances communicate with each other and process data in parallel to create an aggregate output.
These types of scaled-out applications are extremely vulnerable to application breaches, data thefts from cache and memory by scraping, and other methods of illicitly obtaining data from the applications, cache, and/or memory. In data centers which cater to important applications and data types, such as Personally Identifiable Information (PII), Payment Card Industry (PCI) data, medical information that falls under Health Insurance Portability and Accountability Act (HIPAA), military and Government critical tasks, any application and/or data breach is very destructive and expensive to contain and/or resolve. Therefore, it is beneficial to attempt to prevent such breaches.
For systems like data centers and data center applications operating therein, enterprise applications, etc., data encryption and decryption is performed using secure keys. However, secure key generation is a challenge, and is subject to its own breaches and snooping. Protocols like Secure Sockets Layer (SSL), Internet Protocol Security (IPSEC), etc., perform complete packet encryption before sending data across the Internet Protocol (IP) layer.